Xref: utzoo gnu.g++.help:681 comp.lang.c++:12639
Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!ucbvax!tut.cis.ohio-state.edu!unreplyable!garbage
From: jbm%speedy.UUCP@LARRY.MCRCIM.MCGILL.EDU (John B. McCluskey)
Newsgroups: gnu.g++.help,comp.lang.c++
Subject: Please suscribe.
Message-ID: <27f8f603@speedy.UUCP>
Date: 2 Apr 91 20:22:11 GMT
Sender: daemon@tut.cis.ohio-state.edu
Followup-To: gnu.g++.help
Organization: Gatewayed from the GNU Project mailing list help-g++@prep.ai.mit.edu
Lines: 19

I am using D.J. Delorie's port of G++ to MS-DOS, and I am having a severe
problem with the G++ front end bombing randomly 30 to 80 percent of the time.
When it bombs, D.J. Delorie's dos extender (GO32), reports exception 2 at
various addresses.  Exception 2 is the NMI on the PC-AT (& clones),
typically generated by a parity error.  This error is quite puzzling, since
it is non-deterministic.  A compile may bomb, and then succeed on the next
attempt, or it may succeed after rebooting, or it may bomb 5 times in a row.
I suspect a subtle hardware interaction, possibly some condition that is
not correctly trapped, probably in the dos-extender code.  D.J. Delorie
has no email address in the documentation, merely a snail mail address.
Can someone advise?   

Thanks,
-- 
John McCluskey

Postal Mail : John McCluskey, 808 de Bienville, Montreal P.Q. H2J 1V1,  CANADA
Telephone   : (514) 527-2315      Call Sign: KB6PZF
EMail (home): jbm%speedy.uucp@larry.mcrcim.mcgill.edu